Product · 2019 — 2025
Zalo Cloud & Local Storage — Reliable storage with AI-assisted search and auto-categorization
- Role
- Product Manager, Zalo
- Year
- 2019 — 2025
- Team
- Designers, mobile + server engineers, ML/data, QC, research
- Timeframe
- ~6 years, multiple parallel scopes
Summary
Reshaped how people store, back up, and find their stuff in Zalo — a personal cloud, an all-new backup, and a faster local storage manager — built across many concurrent releases with design, client, server, ML and QC.
Problem
Years of media and documents pile up inside chats. People lose receipts, forget where a backup lives, and run out of phone storage with no clear way to clean up. The model also had to be honest about what sits on the phone, what sits in the cloud, and what is actually safe, and it had to keep moving while the rest of the app shipped.
Approach
Treat storage as three connected products running in parallel: a lightweight personal cloud, a real backup, and a local manager. The same designers, client and server engineers, ML/data and QC moved between them as the work demanded, carrying the context with them. Use AI where it removes work from the person, ship in small slices, learn from how devices actually behave in the field.
Outcomes
My Documents
Personal cloud · formerly My Cloud
zCloud
All-new Zalo backup
Local manager
Faster, with smart cleanup suggestions
Process
- 01
My Documents
Reframed the old 'send message to me' habit into a lightweight personal cloud — designed and shipped with the messaging design team, server, and ML for search and auto-categorisation.
- 02
zCloud
An all-new backup for Zalo, built end-to-end with server and client engineers, QC and data — many of them the same people who'd just shipped My Documents — protecting conversations and media while doubling as a way to free up device space.
- 03
Local storage
A faster storage manager with cleanup suggestions, reworked release by release with mobile engineers and data so people could reclaim gigabytes without fearing they'd delete something important.
"The AI that mattered here did an unglamorous job — remembering where something is so nobody has to go looking for it. Six years, one small team across three products, and the point of all of it was that people stop thinking about storage."